Re: Star Schema from 'Flattened' CSV

From: Geoff Muldoon <geoff.muldoon_at_trap.gmail.com>
Date: Fri, 3 Nov 2017 10:03:48 +1100
Message-ID: <MPG.34664f5f61ba689d9896d8_at_news.albasani.net>


In article <3d31cbb6-8546-40fc-9493-dcb47287e727_at_googlegroups.com>, transreductionist_at_gmail.com says...
>
> Clarification from the author:
>
> The first Star Schema is correct I believe. There is one product as it relates to one vendor's unit price. It meets the criteria for 1st normal form.
>
> I understand that the CSV provided does not meet the criteria for 1st normal form, and this is of course the issue. From what I have read, one solution is to create a fact table for each vendor. Is this the best way to handle this?
>
> As a note, there are 4 vendors, and additionally columns for 8 quoting companies as well, which I did not mention for simplicity sake. This would result in 32 normalized fact tables.

32 fact tables, don't go there!

You just have two different facts, quotes and prices. Technically the prices one is probably a "factless fact".

GM Received on Fri Nov 03 2017 - 00:03:48 CET

Original text of this message